Quote Originally Posted by ajaidev View Post
But if you compare a OCed i5 750 to a i7 860 one can notice how well HT does. It does not behave like real cores the added performance varies a lot, in some case HT can boost pretty high numbers and in others the added performance is little or even negative.
But without a 6-core Nehalem not running HT to compare, how can you be sure the performance variations are not simply due to the software? And given that HT in general only benefits between 10-25% in well multi-threaded applications; whose to say it won't be Thuban that suffers more in real world desktop applications which in general aren't that well multi-threaded.